AUTHOR=Li Yi , Wu Wenjing , Sun Yuchen , Yu Dequan , Zhang Yuemei , Wang Long , Wang Yao , Zhang Xiaozhi , Lu Yongkai 

TITLE=The clinical evaluation of atlas-based auto-segmentation for automatic contouring during cervical cancer radiotherapy

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.945053

DOI=10.3389/fonc.2022.945053

ISSN=2234-943X

ABSTRACT=Purpose: To investigate the influence of atlas library size and CT cross-slice number on accuracy and efficiency of atlas-based auto-segmentation (ABAS) method for the automatic contouring of clinical treatment volume (CTV) and organs at risk (OARs) during cervical cancer radiotherapy. 
Methods: Of 140 cervical cancer patients, contours from 20, 40, 60, 80, 100, 120 patients were selected incrementally to create six atlas library groups in ABAS. Another tested 20 patients were automatic contoured with ABAS method and manual contoured by same professional oncologist. Contours include CTV, bladder, rectum, femoral head-L, femoral head-R and spinal cord. CT cross-slice number of tested 20 patients include 61, 65, 72, 75, 81 and 84. Index of dice similarity coefficients (DSC) and Hausdorff Distance (HD) had been used to assess the consistency between ABAS automatic contouring and manual contouring. The randomized blocks analysis of variance and pair-t test had been used for statistical analysis. 
Results: The mean DSC value of “CTV, bladder, femoral head and spinal cord” were all larger than 0.8. Femoral head and spinal cord showed a high degree of agreement between ABAS automatic contouring and manual contouring, with a mean DC > 0.80 and HD < 1 cm in all atlas library groups. Post hoc-Least Significant Difference comparison indicated that no significant difference had been found between different atlas library size with DSC and HD values. For ABAS efficiency, the atlas library size had no effect on the time of ABAS automatic contouring. Time of automatic contouring increased slightly with CT cross-slice number increased, which were 99.9s, 106.8s, 114.0s, 120.6s, 127.9s, 134.8s with CT cross-slice of 61, 65, 72, 75, 81, 84 respectively. 
Conclusion: 20 atlas library sizes and minimum CT cross-slice number included CTV and OARs are enough for insuring accuracy and efficiency of ABAS automatic contouring during cervical cancer radiotherapy.